Informatics Tools for Tumor Heterogeneity in Multiplexed Fluorescence Images Comprehensive genetic profiling has revealed intrinsic molecular variability, or intra-tumor heterogeneity (ITH), in multiple cancers. Heterogeneity is rooted in both genetic and non-genetic factors and evolves within the context of a tumor microenvironment (TME). Not surprisingly, genetic, phenotypic, and TME heterogeneity present major obstacles to optimal cancer diagnosis and treatment; however, the importance of spatial patterning in ITH has been largely overlooked. The spatial distribution of heterogeneity can be critically analyzed with imaging of tissue sections or tumor microarrays (TMAs) using methods such as immunofluorescence (IF) for proteins and fluorescence in situ hybridization (FISH) for DNA and RNA. These fluorescence imaging techniques probe the tumor and surrounding tissue for the expression of proteins, DNA, and RNA in the context of individual cells, sub-cellular domains and clusters of cells within tissue sections. Typically IF/FISH has been restricted to no more than 4-7 proteins/nucleic acids labeled per slide (multiplexed), but new technological advances now allow up to 60 proteins and a few RNA or DNA probes to be labeled on the same multicellular tissue section of up to ca.10 mm (hyperplexed). Larger tumor domains can be analyzed by stitching together images from tissue sections taken from adjacent regions of the tumor. However, the ability to analyze spatial relationships between proteins and nucleic acids at this scale raises several new informatics challenges, such as how to quantitate and characterize spatial ITH and how to interpret ITH data. To address these challenges, a collaborative team of computational biologists, cancer biologists and pathologists at the University of Pittsburgh and engineers and computer scientists at the General Electric Global Research Center (GRC) will develop software for use by cancer biologists and clinicians to quantitate, interpret and visualize spatial ITH in the context of their particular application and s a first step toward constructing diagnostics based on both cancer biomarker expression levels and spatial relationships between cancer and stromal cells.